Process Synchronization in Computer Networks

Abstract

Synchronization is used in real life so as to organize the work to grantee its continuation and preventing what we call blocking, which means no one go on working. Synchronization has a technical meaning in computer field that is related to this meaning but somewhat different. However, synchronization is increasingly used and being an important issue with the development of operating systems which improve the possibility for processes to cooperate with each other even in distributed systems. Processes can operate within one machine using shared memory or through multiple machines using message passing. This thesis searches how to achieve process synchronization either on a single machine or multicomputer systems. In the former one, synchronization can be achieved using multiple methods such as semaphore and monitor, where as in the later one we can use centralized, distributed or token ring algorithms. Then we focus on one of synchronization problems, deadlock. Deadlock is a situation where two or more processes are all blocked and none of them can become unblocked until one of others become unblocked. Three methods for handling deadlock situation: prevention avoidance, and detection. A C++ program has been designed using Message-Passing Interface (MPI) under LINUX operating system to execute a producer – consumer synchronization problem.